Recently, Gamba et al asserted that aspirin consumption leads to a “21% lower risk of melanoma.” This report received widespread international press coverage, and has significant public health implications. The authors reported that 15,089 aspirin users were examined. A total of 9188 nonaspirin, nonsteroidal antiinflammatory drug (NSAID) users were examined, and 35,529 nonusers were examined for a total of 59,806 individuals. Follow-up was for 11.3 years to 11.9 years (median). The populations were poorly matched, and differed significantly in nearly every respect except, interestingly, “history of melanoma.”
The incidence of melanoma in the NSAID users was 89 per 9188, was 115 per 15,089 in aspirin users, and was 344 per 35,529 in the nonusers. The total was 548 per 59,806. This means that the expected incidence of melanoma in any population would be 548 per 59,806, or 0.0092 (approximately 1 case per 110 persons). The values for the 3 groups were 0.0097 for nonaspirin NSAIDs, 0.0076 for aspirin users, and 0.0097 in the NSAID group. Is the claim that melanoma is less common in the group using aspirin sustainable? A chi-square analysis with 2 degrees of freedom provides a value of 5.191 (Graphpad; P = .075, which is not significant). On the basis of these data, I contend that the findings clearly demonstrate that aspirin is not indicated for the primary prevention of melanoma, at least in the populations studied. The finding is important because aspirin use, even in small doses, is not without side effects. Perhaps a prospective study in patients using aspirin for other reasons (eg, atrial fibrillation) would enable this hypothesis to be tested more accurately and precisely.